Pooled analysis of recent studies of magnetic fields and childhood leukemia

Background: Over forty epidemiologic studies have addressed an association between measured or calculated extremely-low-frequency magnetic fields (MF) and childhood leukemia. These studies have been aggregated in a series of pooled analyses, but it has been 10 years since the last such. Methods: We present a pooled analysis combining individual-level data (24,994 cases, 30,769 controls) from four recent studies on MF and childhood leukemia. Results: Unlike previous pooled analyses, we found no increased risk of leukemia among children exposed to greater MF: odds ratio (OR) = 1.01, for exposure ≥0.4 μT (μT) compared with exposures <0.1 μT. Similarly, no association was observed in the subset of acute lymphoblastic leukemia, birth homes, studies using calculated fields, or when geocoding accuracy was ignored. In these studies, there is a decline in risk over time, also evident when we compare three pooled analyses. A meta-analysis of the three pooled analyses overall presents an OR of 1.45 (95% CI: 0.95–2.20) for exposures ≥0.4 μT. Conclusions: Our results are not in line with previous pooled analysis and show a decrease in effect to no association between MF and childhood leukemia. This could be due to methodological issues, random chance, or a true finding of disappearing effect.